In this study, we use symbolic interaction theory, a broad sample including men and women representing a full range of social backgrounds, and a longitudinal design to provide both a frame of reference and a data set for a better understanding of the transition into retirement and to link adjustment to retirement to the broader processes of later life socialization. The proposed data set will enable us to explore four sets of important theoretical and substantive issues: (1) the contrasting predictions of identity continuity, identity crisis, and role realignment theories on the effects of retirement; (2) the activity and social support theory hypotheses that social interaction may reduce some of the potential disruptive effects of retirement; (3) the long neglected questions of whether older females are less satisfied than men in their retirement; and (4) Atchley's (1976) suggestion that there may be temporal variations in adjustment to retirement. A sample of 800 metropolitan Atlanta men and women between the ages of 58 and 64 enable a longitudinal analysis of the influences of pre-retirement social background characteristics, attitudes, and behaviors on post-retirement mental health and involvements. In addition, we propose to interview respondents in the first month of their retirement and then again 6 months, 12 months, and 24 months after retirement in order to better understand the process by which individuals adjust over time to retirement. The data set is special, in that it focuses on multiple roles (not just the work role), contains a rich and diverse set of self variables (role/identity, social commitment, psychological commitment, salience, self-esteem) and role involvements (attention, energy, time, and subjective time), subjective meanings (health, retirement, income adequacy, and judgments about work), as well as a comprehensive set of social network, activity, and social background measures. LISREL will be used to develop structural equations and empirically test the hypotheses and for the effects of a wide range of statistical interactions. The study will provide a clearer picture of the social psychological processes and consequences of retirement which is valuable not only to social research and theory but can be directly applied to counseling programs to help retired and older people maintain productive and meaningful lives.